BIM model components retrieval method based on visual attention

Retrieval is an important way to achieve the interactive visualization of multidimensional dynamic building big data. However, the traditional keyword-based retrieval method lacks the ability to retrieve and rank the BIM model components. In this paper, a new method is proposed for BIM model components retrieval and ranking based on user’s visual attention. Firstly, a visual attention point is picked up by analyzing the interaction principle of users. Then the spatial coordinates of components are extracted on the basis of the IFC definition method. Finally, according to the defined building model indoor road network, the reachable distances between each components and the user’s visual attention point are obtained, and then the optimal components retrieval result is determined by the distances. The experimental result shows that this method not only allows users to quickly, timely, and fully access the required component information, but also satisfies the application of the building model at various stages. © 2018 Computer Society of the Republic of China. All rights reserved.